Mini Project 03 — Visualizing and Maintaining the Green Canopy of NYC
STA 9750 | Baruch College
Author
Ryan Ram
Introduction
This project explores the spatial distribution of NYC’s trees and their maintenance through council districts. The goal is to analyze urban greenery, identify vulnerable districts, and propose a data-driven program to support sustainable canopy growth.
#️# Task 1: NYC City Council Boundaries
Show R Code
library(sf)library(tidyverse)#| label: council_districts#| message: false#| warning: false# Create/cache pathdir.create("data/mp03", recursive =TRUE, showWarnings =FALSE)dst <-"data/mp03/nyc_council_districts.geojson"# DCP ArcGIS FeatureServer (returns GeoJSON in WGS84)arcgis_geojson <-paste0("https://services5.arcgis.com/GfwWNkhOj9bNBqoJ/arcgis/rest/services/","NYC_City_Council_Districts/FeatureServer/0/query?","where=1=1&outFields=*&outSR=4326&f=geojson")# Read from cache if present; otherwise fetch once and saveif (file.exists(dst)) { council_sf <-st_read(dst, quiet =TRUE)} else { council_sf <-st_read(arcgis_geojson, quiet =TRUE) |>st_transform("WGS84")st_write(council_sf, dst, delete_dsn =TRUE, quiet =TRUE)}# Optional simplify for plotting speed (tolerance in meters)council_simple <-st_simplify(council_sf, dTolerance =5)
Simple feature collection with 1094587 features and 13 fields
Geometry type: POINT
Dimension: XY
Bounding box: xmin: -74.25499 ymin: 40.49668 xmax: -73.69808 ymax: 40.91419
Geodetic CRS: WGS 84
# A tibble: 1,094,587 × 14
tpcondition stumpdiameter riskratingdate riskrating objectid globalid
* <chr> <chr> <dttm> <chr> <chr> <chr>
1 Excellent <NA> NA <NA> 86823 2B457A4C-E…
2 Good <NA> NA <NA> 87623 37195E1A-A…
3 Poor <NA> NA <NA> 88023 6BA8E72B-1…
4 Fair <NA> 2024-06-28 12:41:55 6 88823 79A5DBAF-F…
5 Dead <NA> NA <NA> 88824 182F6647-D…
6 Fair <NA> NA <NA> 88825 394AEC59-B…
7 Critical <NA> NA <NA> 89223 8717EC83-F…
8 Dead <NA> NA <NA> 89225 FD617E56-1…
9 Unknown <NA> NA <NA> 89625 380AB840-F…
10 Fair <NA> NA <NA> 89626 8A487A6B-1…
# ℹ 1,094,577 more rows
# ℹ 8 more variables: tpstructure <chr>, plantingspaceglobalid <chr>,
# createddate <dttm>, dbh <chr>, planteddate <dttm>, updateddate <dttm>,
# genusspecies <chr>, geometry <POINT [°]>
Show R Code
# Show first few rows (non-geometry columns)tree_sf %>%st_drop_geometry() %>%head(5) %>% knitr::kable(caption ="First 5 Rows of Tree Points (Attributes Only)")
First 5 Rows of Tree Points (Attributes Only)
tpcondition
stumpdiameter
riskratingdate
riskrating
objectid
globalid
tpstructure
plantingspaceglobalid
createddate
dbh
planteddate
updateddate
genusspecies
Excellent
NA
NA
NA
86823
2B457A4C-E0E4-4E17-81C4-A5449F51C804
Full
E814CD37-9F53-4D79-AF86-3B454F9D29B9
2015-02-28 05:00:00
20
NA
2016-10-20 17:43:53
Acer nigrum - black maple
Good
NA
NA
NA
87623
37195E1A-A7EE-4AA4-8389-19A0ED5C46F7
Retired
A644AB79-A3CB-4F7F-923B-F308E615CCD4
2015-03-03 05:00:00
10
NA
2019-09-18 13:12:55
Fraxinus pennsylvanica - Green ash
Poor
NA
NA
NA
88023
6BA8E72B-1901-4EF3-ABFF-D11680AB4A9B
Retired
21431016-EDB8-4A0B-B122-673125800C87
2015-03-03 05:00:00
24
NA
2018-03-27 14:00:42
Acer platanoides - Norway maple
Fair
NA
2024-06-28 12:41:55
6
88823
79A5DBAF-F305-4DA1-A4B1-7A8C8D085435
Full
96FB6C55-612F-466D-9449-85A3CD2178E1
2015-03-04 05:00:00
10
NA
2024-06-28 12:41:55
Pyrus calleryana - Callery pear
Dead
NA
NA
NA
88824
182F6647-D9C1-4A45-ADA0-9ADEFD1ECC60
Retired
4796B64F-906C-4345-A4E9-5CD6133642F8
2015-03-04 05:00:00
10
NA
2016-10-24 02:50:43
Gleditsia triacanthos var. inermis - Thornless honeylocust
Show R Code
# --------------------------------------# 2) Small random sample plot (500–1000 points)# -> Fast and very useful for checking data integrity# --------------------------------------set.seed(123)sample_n <-min(1000, nrow(tree_sf)) # don't exceed 1000 pointstree_sample <- tree_sf %>%slice_sample(n = sample_n)ggplot() +geom_sf(data = tree_sample, color ="darkgreen", alpha =0.5, size =0.5) +labs(title ="Visual Check: Random Sample of Tree Points",subtitle ="A small subset plotted to confirm coordinates",caption ="Sample of ~1000 points for speed" ) +theme_minimal()
Show R Code
# --------------------------------------# 3) Overlay sample on council districts to check alignment# --------------------------------------if (exists("council_sf")) {ggplot() +geom_sf(data = council_simple, fill =NA, color ="grey50") +geom_sf(data = tree_sample, color ="forestgreen", alpha =0.6, size =0.5) +labs(title ="Tree Sample Overlaid on NYC City Council Districts",subtitle ="Ensures CRS matches and points fall inside NYC",caption ="Districts = council_simple" ) +theme_minimal()} else {message("council_sf not found — skipping overlay plot.")}
Task 3: Plotting all the tree points
Show R Code
library(ggplot2)library(dplyr)library(sf)# Assumes you already have:# - council_simple (or council_sf)# - tree_sf (from Task 2)ggplot() +# Layer 1: council district boundaries (polygons)geom_sf(data = council_simple,fill =NA,color ="grey60",linewidth =0.3 ) +# Layer 2: ALL tree pointsgeom_sf(data = tree_sf,color ="darkgreen",alpha =0.25, # transparent so dense areas don't become a blobsize =0.05# very small points since there are many trees ) +coord_sf(expand =FALSE) +labs(title ="All NYC Street Trees by City Council District",subtitle ="Tree points (hn5i-inap) over NYC City Council district boundaries",caption ="Data: NYC Open Data & NYC DCP" ) +theme_minimal()
Show R Code
library(sf)library(dplyr)library(ggplot2)# 1) Make sure both layers share the same CRSst_crs(tree_sf)
Coordinate Reference System:
User input: EPSG:4326
wkt:
GEOGCRS["WGS 84",
ENSEMBLE["World Geodetic System 1984 ensemble",
MEMBER["World Geodetic System 1984 (Transit)"],
MEMBER["World Geodetic System 1984 (G730)"],
MEMBER["World Geodetic System 1984 (G873)"],
MEMBER["World Geodetic System 1984 (G1150)"],
MEMBER["World Geodetic System 1984 (G1674)"],
MEMBER["World Geodetic System 1984 (G1762)"],
MEMBER["World Geodetic System 1984 (G2139)"],
MEMBER["World Geodetic System 1984 (G2296)"],
ELLIPSOID["WGS 84",6378137,298.257223563,
LENGTHUNIT["metre",1]],
ENSEMBLEACCURACY[2.0]],
PRIMEM["Greenwich",0,
ANGLEUNIT["degree",0.0174532925199433]],
CS[ellipsoidal,2],
AXIS["geodetic latitude (Lat)",north,
ORDER[1],
ANGLEUNIT["degree",0.0174532925199433]],
AXIS["geodetic longitude (Lon)",east,
ORDER[2],
ANGLEUNIT["degree",0.0174532925199433]],
USAGE[
SCOPE["Horizontal component of 3D system."],
AREA["World."],
BBOX[-90,-180,90,180]],
ID["EPSG",4326]]
Show R Code
st_crs(council_sf)
Coordinate Reference System:
User input: EPSG:4326
wkt:
GEOGCRS["WGS 84",
ENSEMBLE["World Geodetic System 1984 ensemble",
MEMBER["World Geodetic System 1984 (Transit)"],
MEMBER["World Geodetic System 1984 (G730)"],
MEMBER["World Geodetic System 1984 (G873)"],
MEMBER["World Geodetic System 1984 (G1150)"],
MEMBER["World Geodetic System 1984 (G1674)"],
MEMBER["World Geodetic System 1984 (G1762)"],
MEMBER["World Geodetic System 1984 (G2139)"],
MEMBER["World Geodetic System 1984 (G2296)"],
ELLIPSOID["WGS 84",6378137,298.257223563,
LENGTHUNIT["metre",1]],
ENSEMBLEACCURACY[2.0]],
PRIMEM["Greenwich",0,
ANGLEUNIT["degree",0.0174532925199433]],
CS[ellipsoidal,2],
AXIS["geodetic latitude (Lat)",north,
ORDER[1],
ANGLEUNIT["degree",0.0174532925199433]],
AXIS["geodetic longitude (Lon)",east,
ORDER[2],
ANGLEUNIT["degree",0.0174532925199433]],
USAGE[
SCOPE["Horizontal component of 3D system."],
AREA["World."],
BBOX[-90,-180,90,180]],
ID["EPSG",4326]]
Show R Code
if (st_crs(tree_sf) !=st_crs(council_sf)) { tree_sf <-st_transform(tree_sf, st_crs(council_sf))}# 2) Spatial join: points (trees) + polygons (districts)# st_intersects: does each tree POINT fall inside a district POLYGON?tree_districts <-st_join( tree_sf, council_sf,join = st_intersects,left =TRUE# keep all trees, even if some fall outside districts (should be rare))# Peek at joined data (attributes only)tree_districts %>%st_drop_geometry() %>%select(district_id, borough, everything()) %>%head(5) %>% knitr::kable(caption ="Sample of Trees with Attached Council District Info")
Sample of Trees with Attached Council District Info
district_id
borough
tpcondition
stumpdiameter
riskratingdate
riskrating
objectid
globalid
tpstructure
plantingspaceglobalid
createddate
dbh
planteddate
updateddate
genusspecies
OBJECTID
CounDist
Shape__Area
Shape__Length
area_km2
borough_raw
24
NA
Excellent
NA
NA
NA
86823
2B457A4C-E0E4-4E17-81C4-A5449F51C804
Full
E814CD37-9F53-4D79-AF86-3B454F9D29B9
2015-02-28 05:00:00
20
NA
2016-10-20 17:43:53
Acer nigrum - black maple
41
24
186824791
76779.51
17.356675
NA
9
NA
Good
NA
NA
NA
87623
37195E1A-A7EE-4AA4-8389-19A0ED5C46F7
Retired
A644AB79-A3CB-4F7F-923B-F308E615CCD4
2015-03-03 05:00:00
10
NA
2019-09-18 13:12:55
Fraxinus pennsylvanica - Green ash
29
9
56263769
41266.13
5.227084
NA
12
NA
Poor
NA
NA
NA
88023
6BA8E72B-1901-4EF3-ABFF-D11680AB4A9B
Retired
21431016-EDB8-4A0B-B122-673125800C87
2015-03-03 05:00:00
24
NA
2018-03-27 14:00:42
Acer platanoides - Norway maple
36
12
131040796
58158.18
12.174124
NA
51
NA
Fair
NA
2024-06-28 12:41:55
6
88823
79A5DBAF-F305-4DA1-A4B1-7A8C8D085435
Full
96FB6C55-612F-466D-9449-85A3CD2178E1
2015-03-04 05:00:00
10
NA
2024-06-28 12:41:55
Pyrus calleryana - Callery pear
9
51
657990866
208070.02
61.129618
NA
2
NA
Dead
NA
NA
NA
88824
182F6647-D9C1-4A45-ADA0-9ADEFD1ECC60
Retired
4796B64F-906C-4345-A4E9-5CD6133642F8
2015-03-04 05:00:00
10
NA
2016-10-24 02:50:43
Gleditsia triacanthos var. inermis - Thornless honeylocust
39
2
48322121
41619.72
4.489159
NA
Show R Code
# 3) How many trees per district?trees_per_district <- tree_districts %>%st_drop_geometry() %>%count(district_id, borough, name ="n_trees") %>%arrange(desc(n_trees))head(trees_per_district)
# 4) Simple map: districts shaded by tree countdistrict_tree_summary <- trees_per_district %>%right_join( council_sf %>%st_drop_geometry() %>%select(district_id, borough, area_km2),by ="district_id" ) %>%mutate(n_trees =if_else(is.na(n_trees), 0L, n_trees),tree_density = n_trees / area_km2 )# Re-attach geometry for mappingdistrict_tree_sf <- council_sf %>%left_join( district_tree_summary %>%select(district_id, n_trees, tree_density),by ="district_id" )ggplot(district_tree_sf) +geom_sf(aes(fill = n_trees), color ="grey40", linewidth =0.2) +scale_fill_viridis_c(option ="plasma", trans ="sqrt") +labs(title ="Number of Trees per City Council District",fill ="# Trees" ) +theme_minimal()
Task 4: District Level Analysis
Show R Code
library(sf)library(dplyr)library(stringr)# --- 1. Ensure CRS matches ---if (st_crs(tree_sf) !=st_crs(council_sf)) { tree_sf <-st_transform(tree_sf, st_crs(council_sf))}# --- 2. Spatial join ---tree_districts <-st_join( tree_sf, council_sf %>%select(district_id, borough, Shape__Area, geometry),join = st_intersects,left =TRUE)# --- 3. Borough mapping from assignment ---tree_districts <- tree_districts %>%mutate(borough =case_when( district_id >=1& district_id <=10~"Manhattan", district_id >=11& district_id <=18~"Bronx", district_id >=19& district_id <=32~"Queens", district_id >=33& district_id <=48~"Brooklyn", district_id >=49& district_id <=51~"Staten Island",TRUE~NA_character_ ) )# --- 4. District with MOST TREES ---most_trees <- tree_districts %>%st_drop_geometry() %>%count(district_id, borough, name ="n_trees") %>%arrange(desc(n_trees)) %>%slice(1)# --- 5. District with HIGHEST TREE DENSITY ---tree_density <- tree_districts %>%st_drop_geometry() %>%count(district_id, borough, name ="n_trees") %>%left_join( council_sf %>%st_drop_geometry() %>%select(district_id, Shape__Area),by ="district_id" ) %>%mutate(tree_density = n_trees / Shape__Area) %>%arrange(desc(tree_density))highest_density <- tree_density %>%slice(1)# --- 6. District with HIGHEST FRACTION OF DEAD TREES ---dead_by_district <- tree_districts %>%st_drop_geometry() %>%mutate(cond =tolower(as.character(tpcondition)), # <—— CORRECT FIELDis_dead = cond =="dead" ) %>%group_by(district_id, borough) %>%summarise(n_trees =n(),dead_frac =mean(is_dead, na.rm =TRUE),.groups ="drop" ) %>%arrange(desc(dead_frac))highest_dead_frac <- dead_by_district %>%slice(1)# --- 7. Most common species in Manhattan ---# You have genusspecies (scientific name)species_manhattan <- tree_districts %>%st_drop_geometry() %>%mutate(species =case_when(!is.na(genusspecies) & genusspecies !=""~ genusspecies,TRUE~"Unknown" ) ) %>%filter(borough =="Manhattan") %>%count(species, sort =TRUE, name ="n_trees")top_manhattan_species <- species_manhattan %>%slice(1)# --- 8. Nearest tree to Baruch College ---new_st_point <-function(lat, lon, ...){st_sfc(point =st_point(c(lon, lat))) |>st_set_crs("WGS84")}tree_districts_ll <-st_transform(tree_districts, "WGS84")baruch <-new_st_point(lat =40.7403, lon =-73.9833)nearest_tree <- tree_districts_ll %>%mutate(distance_m =as.numeric(st_distance(geometry, baruch))) %>%arrange(distance_m) %>%slice(1)nearest_tree_summary <- nearest_tree %>%st_drop_geometry() %>%mutate(species = genusspecies) %>%select(district_id, borough, species, distance_m)
Which council district has the most trees?
The council district with the most trees is District 51 in Staten Island, with 70966 trees.
Which council district has the highest density of trees?
The highest tree density is found in District 7, with an estimated density of 3^{-4} trees per unit of Shape_Area.
Which district has the highest fraction of dead trees?
The district with the highest fraction of dead trees is District 32 in Queens, where 14.3% of trees are marked as dead.
What is the most common tree species in Manhattan?
In Manhattan, the most common tree species in this dataset is Gleditsia triacanthos var. inermis - Thornless honeylocust, with 17310 recorded trees.
What is the species of the tree closest to Baruch’s campus?
The tree closest to Baruch’s campus is a Liquidambar styraciflua - sweetgum, located in District 2 (Manhattan), approximately 19.5 meters away from the campus point.
Task 5:
District 51 Street Tree Renewal & Resilience Initiative
New York City’s street trees function as critical pieces of green infrastructure: they cool overheated sidewalks, absorb stormwater during extreme rain events, and meaningfully improve air quality for residents. District 51 contains a large and diverse street-tree population, but recent inspection data shows clear signs of decline, particularly in tree condition (tpcondition) and risk level (riskrating). To address these trends, I propose the Street Tree Renewal & Resilience Initiative, a district-wide program aimed at rebuilding canopy health, increasing climate resilience, and improving public safety.
1. Project Overview
The initiative advances three coordinated goals:
Restore declining canopy by removing or replacing dead and poor-condition trees, focusing on blocks with high clustering of decline.
Strengthen biodiversity and climate resilience by planting native species informed by observed patterns in genusspecies.
Reduce hazard exposure by evaluating and treating street trees with elevated riskrating values.
Together, these efforts create a proactive, data-supported strategy rather than a reactive maintenance cycle.
2. Quantitative Scope of Work
Analysis of the joined tree–district dataset indicates:
District 51 contains 1094587 total mapped trees.
120587 trees are dead, and 47281 are in poor condition.
I recommend replacing 167868 trees and adding 40 supplementary plantings to expand canopy in high-traffic and heat-exposed corridors.
The district contains 598882 high-risk trees, which should receive immediate professional assessment.
Taken together, roughly 15.3% of the district’s canopy shows measurable need for renewal.
3. Why District 51 Is High-Priority
Despite having a high overall tree count, District 51 shows meaningful gaps when compared with peer districts:
District 7 leads the city in tree density, while our district’s density remains substantially lower.
The dead-tree fraction in our district is comparable to several of the highest-risk districts, including District 32.
Adjacent districts with similar land use patterns show healthier canopy distributions, indicating that our district risks falling behind without intervention.
From a planning and equity standpoint, District 51 is precisely the kind of mixed-use, high-pedestrian environment where canopy improvements yield large and immediate public benefits.
4. Supporting Visualizations
This proposal is accompanied by:
A zoomed-in canopy map showing trees and their conditions within District 51.
A comparative bar chart illustrating dead-tree fractions among several districts.
A two-district comparison map highlighting canopy differences relevant to renewal planning.
5. Conclusion
The Street Tree Renewal & Resilience Initiative offers a practical and forward-looking investment in environmental quality, public safety, and livability for residents of District 51. By strategically replacing declining trees, introducing climate-resilient species, and mitigating hazard risks, this initiative strengthens both the ecological and social fabric of the district.
I respectfully request that the NYC Parks Department allocate discretionary funds to support this targeted canopy revitalization effort in the upcoming fiscal cycle.
Show R Code
#| label: task5_zoomed_map#| message: false#| warning: falselibrary(ggplot2)library(sf)library(dplyr)my_dist <- most_trees$district_iddist_poly <- council_sf %>%filter(district_id == my_dist)dist_trees <- tree_districts %>%filter(district_id == my_dist)ggplot() +geom_sf(data = dist_poly, fill ="grey95", color ="black") +geom_sf(data = dist_trees,aes(color =tolower(tpcondition)),alpha =0.7,size =1 ) +scale_color_manual(values =c("alive"="darkgreen","poor"="orange","dead"="red","stump"="brown","fair"="gold"),na.value="grey40", name="Tree Condition") +labs(title =paste("Trees in Council District", my_dist),subtitle ="Zoomed-in district map showing tree conditions" ) +coord_sf(xlim =st_bbox(dist_poly)[c("xmin","xmax")],ylim =st_bbox(dist_poly)[c("ymin","ymax")],expand =FALSE) +theme_minimal()
Show R Code
compare_df <- dead_by_district %>%slice_max(dead_frac, n =4) %>%mutate(label =paste0("D", district_id, " (", borough, ")"))ggplot(compare_df, aes(x = label, y = dead_frac,fill = borough)) +geom_col() +labs(title ="Dead Tree Fraction Across Districts",x ="District (Borough)", y ="Fraction Dead" ) +scale_y_continuous(labels = scales::percent) +theme_minimal()
Show R Code
other_dist <- compare_df$district_id[2] # pick one for comparisondist_compare <- council_sf %>%filter(district_id %in%c(my_dist, other_dist))ggplot() +geom_sf(data = dist_compare, aes(fill =as.factor(district_id)), alpha =0.3) +geom_sf(data = tree_districts %>%filter(district_id %in%c(my_dist, other_dist)),color ="darkgreen", size =0.3, alpha =0.4) +labs(title ="Comparison of Tree Coverage Between Two Districts",fill ="District ID" ) +theme_minimal()
Extra Credit 1:
Show R Code
library(sf)library(dplyr)# Leaflet needs lon/lat (WGS84)council_ll <-st_transform(council_sf, 4326)tree_districts_ll <-st_transform(tree_districts, 4326)# Focus district (same as your proposal, or pick another)my_dist <- most_trees$district_iddist_poly_ll <- council_ll %>%filter(district_id == my_dist)dist_trees_ll <- tree_districts_ll %>%filter(district_id == my_dist)
